部分 Claude 用户不满 Anthropic 新水印会让他们在工作和课堂上使用 AI 时被识别

Anthropic 最近决定为 Claude 的输出添加水印——在聊天机器人的文字输出中嵌入不可见代码，以标记其为 AI 生成内容。

Anthropic 推出这项新政策，是为了满足《欧盟 AI 法案》透明度准则；该准则如今要求科技公司对由 AI 生成或编辑、且可被计算机系统识别的内容进行标注。然而，欧洲监管机构或许会对此感到满意，部分 AI 用户显然并不高兴。

只要看看 Reddit，就能发现正在酝酿的不满情绪，尽管该网站上的其他发帖者并不认同这些看法。我看到的较为情绪化的一篇帖子来自一名名为 visionode 的用户，其账号值得注意的是仅注册了三周。按照 visionode 的说法，这套新水印系统是一场严苛的阴谋，旨在迫害全世界无辜的聊天机器人用户。

Visionode 的基本论点似乎是：熟练的 Claude 用户或许能通过改写文本，或借助其他 AI 服务清理输出内容，来隐藏自己使用 AI 的证据；但普通 Claude 用户将会被抓个正着。

“谁会被抓到？你。那个用 Claude 重新组织一段文字的学生。那个让 AI 总结一份长达两百页记录稿的记者。那个陷入创作瓶颈、请 AI 提供同义词的作者。这些人走完流程后，额头上都会留下一个数字纹身。”

我无意贬低 visionode 的愤怒，但这些并不是最好的例子。让 AI 总结一份两百页记录稿的记者，不会在意该摘要附带水印，除非他们把这份摘要原封不动地复制粘贴进自己的文章——这显然不合伦理，也不应当发生。

同样地，一名学生让 Claude“重新组织一段文字”后，再将其输出复制粘贴进论文，也是不合伦理的行为。

其他 Reddit 用户也并不特别支持原帖作者的愤怒。

“看看这家伙。”一名发帖者只是这样评论。

另一人则要求原帖作者先“深呼吸”。

Visionode 并不是唯一抱怨的人。另一名不满的用户称这些水印“毫无伦理”且“令人厌恶”，并主张自己在使用 Claude 时已经完成了绝大部分工作。在其看来，聊天机器人只是一个帮助其完成艰苦劳动的“工具”。

“我提供了指令、背景、决策以及无数次调整，Claude 是工具。如果 Claude 开始给它生成的代码或其他任何内容加水印，那么它到底是在宣称对什么拥有功劳？”该发帖者问道。

同样，其他用户也纷纷批评这位质疑者。

“它并没有在宣称功劳，”一名用户反驳道，“这是为了能够检测 AI 生成的输出，因为在各种场景中，AI 生成的输出可能带来风险。”

“这哥们儿连抱怨 Claude 都得用 Claude 来写。”另一人调侃道。

一些批评者避开了受害者叙事，对 Anthropic 的新政策提出了稍微更细致的反对理由。

例如，一名发帖者抱怨，对一款本身通过大量吸收他人作品而生成的文字产品添加水印，存在普遍性的虚伪。“我认为这是一个非常阴险的发展方向，”这名用户表示，“我并不使用 Claude 写作，但考虑到许多前沿模型的训练数据来源，让一个会给你的作品加水印的 AI 出现，实在是讽刺得令人害怕。”

不过总体而言，用户倾向于支持水印系统，认为这是追踪算法生成材料的一种合理方式。“根本没有任何站得住脚的理由说明这不是个好主意，”另一条讨论中的一名用户说道，“你唯一不希望这样做的理由，就是想对别人撒谎。”

Anthropic recently made the decision to watermark Claude’s outputs — inserting invisible code into the chatbot’s editorial text that marks it as AI-generated.

Anthropic rolled out this new policy to satisfy the EU AI Act’s Transparency Code, which now requires tech companies to label content that has been AI-generated or edited in a manner identifiable to computer systems. Yet while European regulators may be happy, some AI users are decidedly not.

One need look no further than Reddit to find evidence of brewing discontent, though other posters on the site are not in agreement. One of the more histrionic posts I came across was from a user named visionode, whose account is notably only three weeks old. According to visionode, the new watermarking system is a draconian conspiracy designed to victimize innocent chatbot users worldwide.

Visionode’s basic argument appears to be that, while savvy Claude users may be able hide evidence of their AI usage by paraphrasing or otherwise cleaning their outputs through other AI services, the average user of Claude will be caught.

“Who will get caught? You. The student who used Claude to reorganize a paragraph. The journalist who asked the AI to summarize a two-hundred-page transcript. The writer who had creative block and asked for synonyms. Those guys come out of the process with a digital tattoo on their forehead.”

Far be it from me to undermine visionode’s outrage, but those are not the best examples. A journalist asking AI to summarize a two-hundred-page transcript is not going to be bothered by a watermark attached to that summary, unless they are copying and pasting the summary verbatim into their article — which is plainly unethical and shouldn’t be happening.

It is equally unethical for a student who copies and pastes Claude’s output into an essay after asking it to “reorganize a paragraph.”

Other Redditors were also not particularly supportive of the poster’s outrage.

“Get a load of this guy,” one poster merely commented.

Another asked the OP to take “a deep breath.”

Visionode wasn’t the only one complaining. Another unhappy camper called the watermarks “unethical” and “disgusting” and argued that by using Claude, they had done the lion’s share of the work. In their view, the chatbot was merely a “tool” that had facilitated their arduous labor.

“I gave the instructions, context, decisions, and countless refinements, claude was the tool. If Claude starts watermarking the code or anything else it generates, what exactly is it claiming credit for?” the poster asked.

Again, other users dogpiled onto the critic.

“It’s not claiming credit though,” one user shot back. “It’s about being able to detect AI generated outputs because of the risks AI generated outputs can cause in various situations.”

“Bro couldn’t even complain about Claude without using Claude to write it,” another quipped.

Some critics have steered clear of the victimhood narrative and made slightly more nuanced arguments against Anthropic’s new policy.

For instance, one poster complained of a general hypocrisy in watermarking an editorial product that was, itself, generated by hoovering up other people’s work. “I think it’s a very sinister direction to take,” said the user. “I don’t use Claude to write anything but having an AI that watermarks your work is terrifyingly ironic given how many of the frontier models came by their training data.”

In general, however, users have tended to support the watermarking system as a sensible way to track material that was generated by algorithm. “There is literally no good argument for why this isn’t a good idea,” a user on another thread said. “The only reason you wouldn’t want this is to lie to people.”